general state of and roadmap for the library?

[KonradHoeffner]

horned-owl is exactly what we need but my colleague tried it out last year and said it was in a really early state.
However seeing that there is a paper about it now which says "The library itself is now mature" and the version is 1.0.0, I thought it would now be ready, but when trying it out I found it really hard to find examples and documentation and I get lots of errors (e.g. a file that exists is "not found", loading an RDF/XML file results in "not implemented") and the last release is 9 months old.
Should we be using the devel branch directly from GitHub instead of the 1.0.0 release?
And what is the roadmap for the project? Is it still actively developed and open to contributions?
If yes, I will put more time into understanding it and I would be happy to offer to help out with adding documentation and examples and so on, however I first wanted to make sure it is still active.

[filippodebortoli]

@phillord could you reply to this?

I would be interested in maintaining the crate more actively as well (when time permits), since there are several improvements to be done across all modules, most importantly fixing the RDF/XML (see #124 #123 for the latest issues), as well as adding a mostly-needed documentation.

My impression is that contributions are welcome if they target the issues raised by @phillord - I had PRs that were accepted and PRs that were rejected even though they were bringing some improvements to the crate.

Therefore, I would also like to ask to @phillord if a) new contributions are welcome, b) if there is a clear roadmap for the next release or c) if you are not planning to actively maintain the crate and if so if d) we can find a way to maintain the crate so that someone else can actively work on it for the next releases?

[phillord]

@KonradHoeffner Well, Horned-OWL is a slow moving piece of software. During it's lifetime it has been an academic piece of software in the sense that no one has even been directly funded to work on it. The software is mature in the sense that it is now a OWL2 complete implementation. The core library is, I think, pretty much feature complete; I would expect that any new big chunks of functionality would be split into other crates (which is maybe where the IO functionality should be).

But it is not hardened in the sense that it hasn't been tested against all ontologies and, joy of the OWL2 spec which is a complicated beast means that when we do try new ontologies we do see bugs. Or in some cases, areas where the specification is ambiguous and I haven't decided what to do yet.

The inline documentation is not too bad, but yes, I agree there is no easy to use tutorial documentation.

In terms of roadmap, tutorial documentation, removing the todo and reconsidering error handling in general, would be high on the list. After that, it would be trying it religiously against lots of ontologies.

Yes, it is open to be contribution -- examples and documentation would be more than welcome. We have a loose development process: do a PR against devel. Releases will happen periodically when there is a new feature or just a collection of existing ones (and as you note, a 1.1 release is probably due).

[phillord]

@filippodebortoli PRs are welcome and you have submitted quite a number. If I remember correctly, I only turned down on PR and it was not because it wasn't targetting an issue raised by me, but because I was worried that it introduced too much complexity into the code for unclear gain.

As is the way with this sort of project, I don't think I have a clear plan for what should be in the next release. It will depend on what people contribute. What would you like to see in the next release?